* Re: [PATCH] Bluetooth: Fix debugfs NULL pointer dereference
[not found] <20180928041039.GA18262@makarhum-Latitude-E5440>
@ 2018-09-28 6:52 ` Marcel Holtmann
0 siblings, 0 replies; only message in thread
From: Marcel Holtmann @ 2018-09-28 6:52 UTC (permalink / raw)
To: Matias Karhumaa; +Cc: Marcel Holtmann, Johan Hedberg, Bluez mailing list
Hi Matias,
> Fix crash caused by NULL pointer dereference when debugfs functions
> le_max_key_read, le_max_key_size_write, le_min_key_size_read or
> le_min_key_size_write and Bluetooth adapter was powered off.
>
> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=201253
>
> BUG: unable to handle kernel NULL pointer dereference at 00000000000002e8
> PGD 0 P4D 0
> Oops: 0000 [#24] SMP PTI
> CPU: 2 PID: 6255 Comm: cat Tainted: G D OE 4.18.9-200.fc28.x86_64 #1
> Hardware name: LENOVO 4286CTO/4286CTO, BIOS 8DET76WW (1.46 ) 06/21/2018
> RIP: 0010:le_max_key_size_read+0x45/0xb0 [bluetooth]
> Code: 00 00 00 48 83 ec 10 65 48 8b 04 25 28 00 00 00 48 89 44 24 08 31 c0 48 8b 87 c8 00 00 00 48 8d 7c 24 04 48 8b 80 48 0a 00 00 <48> 8b 80 e8 02 00 00 0f b6 48 52 e8 fb b6 b3 ed be 04 00 00 00 48
> RSP: 0018:ffffab23c3ff3df0 EFLAGS: 00010246
> RAX: 0000000000000000 RBX: 00007f0b4ca2e000 RCX: ffffab23c3ff3f08
> RDX: ffffffffc0ddb033 RSI: 0000000000000004 RDI: ffffab23c3ff3df4
> RBP: 0000000000020000 R08: 0000000000000000 R09: 0000000000000000
> R10: ffffab23c3ff3ed8 R11: 0000000000000000 R12: ffffab23c3ff3f08
> R13: 00007f0b4ca2e000 R14: 0000000000020000 R15: ffffab23c3ff3f08
> FS: 00007f0b4ca0f540(0000) GS:ffff91bd5e280000(0000) knlGS:0000000000000000
> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00000000000002e8 CR3: 00000000629fa006 CR4: 00000000000606e0
> Call Trace:
> full_proxy_read+0x53/0x80
> __vfs_read+0x36/0x180
> vfs_read+0x8a/0x140
> ksys_read+0x4f/0xb0
> do_syscall_64+0x5b/0x160
> entry_SYSCALL_64_after_hwframe+0x44/0xa9
>
> Signed-off-by: Matias Karhumaa <matias.karhumaa@gmail.com>
> ---
> net/bluetooth/smp.c | 62 ++++++++++++++++++++++++++++++++++++++++++++++++-----
> 1 file changed, 57 insertions(+), 5 deletions(-)
>
> diff --git a/net/bluetooth/smp.c b/net/bluetooth/smp.c
> index 3a7b077..0da3754 100644
> --- a/net/bluetooth/smp.c
> +++ b/net/bluetooth/smp.c
> @@ -3368,9 +3368,22 @@ static ssize_t le_min_key_size_read(struct file *file,
> size_t count, loff_t *ppos)
> {
> struct hci_dev *hdev = file->private_data;
> + struct l2cap_chan *chan;
> + struct smp_dev *smp;
> char buf[4];
>
> - snprintf(buf, sizeof(buf), "%2u\n", SMP_DEV(hdev)->min_key_size);
> + if (!hdev)
> + return -EFAULT;
the -EFAULT is never the right error unless you have some memory access error. And frankly I have no idea what you are checking !hdev here. If that would ever happen, then we have serious other problems.
> +
> + chan = hdev->smp_data;
> + if (!chan)
> + return -EFAULT;
> +
> + smp = chan->data;
> + if (!smp)
> + return -EFAULT;
> +
> + snprintf(buf, sizeof(buf), "%2u\n", smp->min_key_size);
>
> return simple_read_from_buffer(user_buf, count, ppos, buf, strlen(buf));
> }
> @@ -3380,10 +3393,23 @@ static ssize_t le_min_key_size_write(struct file *file,
> size_t count, loff_t *ppos)
> {
> struct hci_dev *hdev = file->private_data;
> + struct l2cap_chan *chan;
> + struct smp_dev *smp;
> char buf[32];
> size_t buf_size = min(count, (sizeof(buf) - 1));
> u8 key_size;
>
> + if (!hdev)
> + return -EFAULT;
> +
> + chan = hdev->smp_data;
> + if (!chan)
> + return -EFAULT;
> +
> + smp = chan->data;
> + if (!smp)
> + return -EFAULT;
> +
> if (copy_from_user(buf, user_buf, buf_size))
> return -EFAULT;
>
> @@ -3391,11 +3417,11 @@ static ssize_t le_min_key_size_write(struct file *file,
>
> sscanf(buf, "%hhu", &key_size);
>
> - if (key_size > SMP_DEV(hdev)->max_key_size ||
> + if (key_size > smp->max_key_size ||
> key_size < SMP_MIN_ENC_KEY_SIZE)
> return -EINVAL;
>
> - SMP_DEV(hdev)->min_key_size = key_size;
> + smp->min_key_size = key_size;
Maybe we just need to store min and max key sizes in hdev instead of trying to store them in the SMP object. Maybe Johan has some other idea, but I think that debugfs command should always work and remember the value even when SMP gets only created later.
Regards
Marcel
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2018-09-28 6:52 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <20180928041039.GA18262@makarhum-Latitude-E5440>
2018-09-28 6:52 ` [PATCH] Bluetooth: Fix debugfs NULL pointer dereference Marcel Holtmann
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).